Content category scoring for nodes in a linked database

• US 8,423,554 B2
• Filed: 07/07/2009
• Issued: 04/16/2013
• Est. Priority Date: 07/07/2009
• Status: Expired due to Fees

• Current Assignee: Sosvia Incorporated
• Original Assignee: Sosvia Incorporated
• Inventors: Garcia, Fermin Otniel, Saldivar, Roberto Romero, Zazueta, Jose Roberto Zavala
• Primary Examiner(s): Perveen, Rehana
• Assistant Examiner(s): HAVAN, HUNG T
• Application Number:
  US12/498,407
• Publication Number:
  US 20110010365A1
• Time in Patent Office: 1,379 Days
• Field of Search: 707/736, 707/738
• US Class Current: 707/738
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/9558 Details of hyperlinks; Management of linked annotations

Abstract

Systems, methods and computer program products are provided for assigning content category scores to nodes of a linked database. The nodes of the linked database include linking nodes and linked nodes. Each linking node is assigned a linking node score for each content category. The linking node score for each content category represents a degree of relevancy of the linking node to the content category. Each of the linked nodes is linked to by at least one of the linking nodes. Each linked node is assigned a content category score for each content category. The content category score for each content category is based on the linking node score for said content category of each linking node that links to the linked node.

Claims

1. An information retrieval system comprising a processor for executing at least one program module for scoring nodes in a linked database based on a plurality of content categories, said nodes including linking nodes and linked nodes, wherein at least one of the linking nodes is a seed node, the seed node having an assigned seed node score for each content category, said at least one program module being stored in a system memory and comprising computer-executable instructions for:

identifying a plurality of linking nodes in the linked database and assigning to each linking node a linking node score for each content category, wherein the linking node score for each content category represents a degree of relevancy of the linking node to the content category;

identifying a plurality of linked nodes in the linked database, the linked nodes being linked to by at least one of the linking nodes;

assigning to each linked node a content category score for each content category, wherein the content category score for each content category is based on the linking node score for said content category of each linking node that links to the linked node;

identifying at least one unique keyword in the seed node;

for each identified unique keyword, determining a keyword score for each content category, wherein the keyword score for each content category is based on the seed node score for said content category divided by the number of unique keywords in the seed node;

determining that a selected content category score for a selected linked node in a selected content category does not exceed a threshold value; and

replacing the selected content category score with a leaf node score that is based on an average of the keyword scores in the selected content category assigned to any of the unique keywords contained in the selected node.

2. The information retrieval system of claim 1, wherein the content category score for each content category is based on the sum of each linking node score for said content category multiplied by a damping factor for said content category, wherein said damping factor reduces the content category score of the linked node based on an estimated rate of irrelevancy of the nodes in the linked database to said content category.

3. The information retrieval system of claim 1, wherein the seed node contains a number of forward links;

and

wherein the linking node score assigned to the seed node for each content category is based on the seed node score assigned to said seed node for said content category divided by the number of forward links in said seed node.

4. The information retrieval system of claim 1, wherein at least one of the linking nodes is a branch node that is linked to by at least one of the other linking nodes;

wherein the branch node contains a number of forward links and has a calculated branch node score for each content category, wherein the branch node score for each content category is based on each linking node score for said content category of the at least one other linking nodes; and

wherein the linking node score assigned to the branch node for each content category is based on the branch node score assigned to said branch node for said content category divided by the number of forward links in said branch node.

5. An information retrieval system comprising a processor for executing at least one program module for scoring nodes in a linked database based on a plurality of content categories, said nodes including linking nodes and linked nodes, wherein a plurality of the linking nodes are seed nodes, each seed node having an assigned seed node score for each content category, wherein the at least one program module is stored in a system memory and comprises computer-executable instructions for:

identifying a plurality of linking nodes in the linked database and assigning to each linking node a linking node score for each content category, wherein the linking node score for each content category represents a degree of relevancy of the linking node to the content category;

identifying a plurality of linked nodes in the linked database, the linked nodes being linked to by at least one of the linking nodes;

assigning to each linked node a content category score for each content category, wherein the content category score for each content category is based on the linking node score for said content category of each linking node that links to the linked node;

identifying at least one unique keyword in each seed node;

for each identified unique keyword in each seed node, determining a keyword score for each content category, wherein the keyword score for each content category is based on the seed node score for said content category divided by the number of unique keywords in the seed node;

determining an aggregate keyword score for each content category by aggregating the keyword scores for each content category of all of the seed nodes;

determining that a selected content category score for a selected linked node in a selected content category does not exceed a threshold value; and

replacing the selected content category score with a leaf node score that is based on an average of the aggregate keyword scores in the selected content category assigned to any of the unique keywords contained in the node.

6. A computer implemented method for scoring a node in a linked database, the node being linked to by a plurality of linking nodes, the method comprising the steps of:

for each linking node, determining a linking node score for each of a plurality of content categories, wherein the linking node score for each content category represents a degree of relevancy of the linking node to the content category;

assigning to the node a content category score for each of the plurality of content categories, wherein the content category score for each content category is based on each linking node score for said content category;

identifying a plurality of unique keyword in a seed node having an assigned seed node score for each content category;

for each identified unique keyword, determining a keyword score for each content category, wherein the keyword score for each content category is based on the seed node score for said content category divided by the number of unique keywords in the seed node; and

if the content category score in any content category does not exceed a threshold value, replacing said content category score with a leaf node score that is based on an average of the keyword scores in said content category assigned to any of the unique keywords contained in the node.

7. The method of claim 6, wherein the content category score for each content category is based on the sum of each linking node score for said content category multiplied by a damping factor for said content category, wherein said damping factor reduces the content category score of the linked node based on an estimated rate of irrelevancy of the nodes in the linked database to said content category.

8. The method of claim 6, wherein the seed node contains a number of forward links;

and

wherein the linking node score assigned to the seed node for each content category is based on the seed node score assigned to said seed node for said content category divided by the number of forward links in said seed node.

9. The method of claim 6, wherein at least one of the linking nodes is a branch node that is linked to by at least one of the other linking nodes;

wherein the branch node contains a number of forward links and has a calculated branch node score for each content category, wherein the branch node score for each content category is based on each linking node score for said content category of the at least one other linking nodes; and

wherein the linking node score assigned to the branch node for each content category is based on the branch node score assigned to said branch node for said content category divided by the number of forward links in said branch node.

10. A computer implemented method for scoring a node in a linked database, the node being linked to by a plurality of linking nodes, wherein a plurality of the linking nodes are seed nodes, each seed node having an assigned seed node score for each content category and wherein the method comprises the steps of:

for each linking node, determining a linking node score for each of a plurality of content categories, wherein the linking node score for each content category represents a degree of relevancy of the linking node to the content category;

assigning to the node a content category score for each of the plurality of content categories, wherein the content category score for each content category is based on each linking node score for said content category

identifying a plurality of keywords in each seed node having an assigned seed node score;

for each identified keyword in each seed node, determining a keyword score for each content category, wherein the keyword score for each content category is based on the seed node score for said content category divided by the number of keywords in the seed node;

determining an aggregate keyword score for each content category by aggregating the keyword scores for each content category of all of the seed nodes; and

if the content category score in any content category of the node does not exceed a threshold value, replacing said content category score with the leaf node score that is based on an average of the aggregate keyword scores in said content category assigned to any of the keywords contained in the node.

11. A computer program product having stored thereon at least one program module for scoring a node in a linked database, the node being linked to by a plurality of linking nodes, the at least one program module comprising computer executable instructions for:

for each linking node, determining a linking node score for each of a plurality of content categories, wherein the linking node score for each content category represents a degree of relevancy of the linking node to the content category;

assigning to the node a content category score for each of the plurality of content categories, wherein the content category score for each content category is based on each linking node score for said content category;

identifying a plurality of unique keyword in a seed node having an assigned seed node score;

for each identified unique keyword, determining a keyword score for each content category, wherein the keyword score for each content category is based on the seed node score for said content category divided by the number of unique keywords in the seed node; and

if the content category score in any content category does not exceed a threshold value, replacing said content category score with a leaf node score that is based on an average of the keyword scores in said content category assigned to any of the unique keywords contained in the node.

12. The computer program product of claim 11, wherein the content category score for each content category is based on the sum of each linking node score for said content category multiplied by a damping factor for said content category, wherein said damping factor reduces the content category score of the linked node based on an estimated rate of irrelevancy of the nodes in the linked database to said content category.

13. The computer program product of claim 11, wherein the seed node contains a number of forward links;

and

wherein the linking node score assigned to the seed node for each content category is based on the seed node score assigned to said seed node for said content category divided by the number of forward links in said seed node.

14. The computer program product of claim 11, wherein at least one of the linking nodes is a branch node that is linked to by at least one of the other linking nodes;

wherein the branch node contains a number of forward links and has a calculated branch node score for each content category, wherein the branch node score for each content category is based on each linking node score for said content category of the at least one other linking nodes; and

wherein the linking node score assigned to the branch node for each content category is based on the branch node score assigned to said branch node for said content category divided by the number of forward links in said branch node.

15. A computer program product having stored thereon at least one program module for scoring a node in a linked database, the node being linked to by a plurality of linking nodes, wherein a plurality of the linking nodes are seed nodes, each seed node having an assigned seed node score for each content category and wherein the at least one program module comprises computer executable instructions for:

for each linking node, determining a linking node score for each of a plurality of content categories, wherein the linking node score for each content category represents a degree of relevancy of the linking node to the content category;

assigning to the node a content category score for each of the plurality of content categories, wherein the content category score for each content category is based on each linking node score for said content category;

identifying a plurality of keywords in each seed node having an assigned seed node score;

for each identified keyword in each seed node, determining a keyword score for each content category, wherein the keyword score for each content category is based on the seed node score for said content category divided by the number of keywords in the seed node;

determining an aggregate keyword score for each content category by aggregating the keyword scores for each content category of all of the seed nodes; and

if the content category score in any content category of the node does not exceed a threshold value, replacing said content category score with the leaf node score that is based on an average of the aggregate keyword scores in said content category assigned to any of the keywords contained in the node.